Author Topic: Keithley 2002 8.5 digit DMM review and teardown  (Read 62339 times)

0 Members and 1 Guest are viewing this topic.

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Keithley 2002 8.5 digit DMM review and teardown
« on: February 18, 2014, 07:24:40 pm »
Finally, it's here. I had no luck to get infamous 3458A affordable enough, so went to very familiar route, with
Keithley. Like some say it can't be enough DMMs, no matter how many one have.
As some members know, i'm Keithley addict in a way, and have few Model 2001 meters, which were violently abused and
tear down to pieces on daily basis, covering journey in this thread at EEVBlog and my own project tracker.

Welcome Model 2002 in it's beauty, probably smallest 8.5 digit meter on market. It's still current model, and
officially avalable thru official manufacturer website, just like 20'year old 3458A :).
To my surprise this one don't require any fixing/repairing or other mess, very unlike my 20 year old 2001's :D

Everybody who saw Keithley 200x/7001/24xx/26xx ones, will know how 2002 look like, so I'll post overall view
and menu/settings tricks later. PDF manuals.

As with all recent Keithley gear, schematics or detailed construction information is not available.
Best we have is repair manual troubleshooting hints, self-test diagnostics description and part list (which is helpful too).

For now we start with real stuff, teardown and

I actually shoot video-review (planned to be short, but it blew to 24 minutes). Sorry, my english require
lot of improvement and it's not my native, and it's almost my first video, so don't judge too harsh :)



And resources + hi-resolution photos of things inside below:
All photos clickable/zoomable :D



Digital board. It's 2001-142, Revision K. Very same board as in my second Keithley 2001.

Differences on BOM level:

* High-voltage DC-DC module (block near front panel connector) (+5V to filament VAC and +60VDC for VFD) moved from digital board to front panel.
* STD option used DALLAS DS1243 (NVRAM with Phantom RTC clock), MEM2 likely to use bigger version, such as DS1248 or DS1251
* Extra I2C 24C16 EEPROM on position U633 for calibration constants/data?
* Firmware for Model 2002
* Missing protection diodes on GPIB port interface (while www.38hot.net BBS 2002 had them on)
* Small ALTERA MAX5064 CPLD for trigger-link replaced by TEKMOS 2001-801-A05 ASIC
* 4MHz clock moved to front panel too.



Analog board is different story tho...



In general construction and functional units are same, but in critical areas using more precise components with tigther tolerances/less tempco.
Voltage reference changed from aged/selected LM199/399 to LTZ1000ACH.
There are now two switching capacitor building blocks (Linear LTC1043CSW) paired with KEMET PC or PP film capacitors.
Current source for resistance is using Vishay foil's. Switching between ranges now using MAX326's.
Ohm measurement is almost same, 0.1ohm 1% kelvin resistor (but better tempco, +-10ppm by label on it) + 0.91ohm and
Caddock 1737-53 network (same as 2001).
Same bootstrap circuitry as 2001's (four big caps near transformer) with avalance diodes between them,
same True-RMS AD637JR.



2002-162 4-layer PCB with ADC module now mostly completely SMT, designed in 1993 (yes, 21 years ago!).
It's revision B.

Featuring ALTERA EPM7160 in PLCC84 package, configured to 2002-802-A01 RTL, and clocked from 12MHz FOX XO.
As reminder, Model 2001 was using 7.68MHz with EPM7128 in PLCC64 :)

Some resistors now are in SO-16 networks (R827 and R825 on left side)
Also I could not spot integrating capacitor, it's probably one of NP0 SMD caps.
Model 2001 ADC was using bulky film thru-hole cap rated to 400V.



No components on back side, only nicely labelled interfacing connector. Reference voltage supplied differentially,
REFLO and REFHI? I measured REFHI, it's 7.13..V

Let's go back to analog board...



Mains switching / control curcuitry is same as in 2001. It handles selection 90-132 or 180-250VAC ranges automatically, as
well as provide frequency zero-cross signal for ADC to trigger in right at required NPLC cycles, to increase CMRR and PSRR.



Optical isolation with digital board was done with separate daughterboard, labelled 2001-132-01A. It carry two NVE IL717, which
are four-channel 110Mbps digital isolators. But if look closely - pin interface between boards matching original TLP582's,
so my guess that TLP582 couples were difficult to get in stock, so they added this adapter board, to keep analog board
design intact.



Power transformer have same configuration and pinout as 2001, marked HIREL systems, TR-280 Revision D.
It have magnetic shielding with copper foil around winding.



2002 equipped with very similar or same bootstrap circuitry, which provides +39 and -39V for analog needs, as well as +8/-8VF
for ohm/amps measurements. Those pesky C116,C115,C118,C119 capacitors tend to leak with age/elevated temperatures and
commit complete disaster on PCB, eating traces, leaking electrolyte all over precision circuits... This meter is only 8 years
old, so they are still fine, but next year I will replace them to freshies for sure :)





Input side.



PCB Fuse :) Old 2002 analog board don't have this, at least from what I can say looking on crappy photo
from chinese forum mentioned before. 2001 also don't have this. Wonder what that one does, maybe some
last step emergency protection?

Combined Caddock resistor network.



Consists from two resistors, bigger TF050R, and glued TF020R on back of it.



Reed relays made by Americal Relays Inc in this meter, A051A300H0C model, which likely to be custom, and
usual DK1a1b-6V relay. Circut here same or almost same as 2001.



Old 8-bit DAC from 2001 meter (MP7528) got replaced by same thing, which is Analog Devices AD7528, dual matched DACs, which
are used for TRMS filter circutiry to adjust operation of meter's AC conversion block. This section is very similar or same to 2001.



Same with TrueRMS convertor itself, which is AD637JR with bunch of DG444,DG408 and DG41 switches around.
There are also AD848J, LT1097, LT1223, AD711J and LM79L05A -5V negative LDO in this area.
All same stuff as 7.5 digit Keithley brother.







7V VREF:



Famous Linear ultra-zener reference, LTZ1000.



It have dedicated plastic cup from both sides of to seal reference from any airflow and reduce temperature influence from rest of the world.
Probably it's reference standard connection, can see 120 ohm 0.1% resistor nearby, as well as few PTF56 low-tempco resistors.
Also two blue resistors networks nearby, but those may not be related to voltage reference, at least one of such networks
can be found in 2001 too.



Back side of PCB with LTZ.



LTZ have popular companion Linear LT1013 in SO8 package with Q204 is heater drive transistor in SMT package.



There is also temperature sensor somewhere on board, as Keithley 2002 can measure it's internal temp, which it reports as
delta to calibration temperature. This measurement is readily available by temperature function, by selection INTERNAL TEMP mode.
In my case, reported value was -3.4°C , with ambient temp +24°C, while meter run around half-hour.
If I just power on cold meter it says -23.2°C and descreasing (to lesser delta) while meter warms up.



Current measurement parts, two MOSFETs to switch ranges for high amp ranges, long Vishay Z201 resistor (R277).
Right side of photo is for resistance measurements.



Same CADDOCK as in 2001 for low current ranges. This is actually standard network 1737-53, available even on Digikey :)



And not so standard Vishay PG hermetic oil-filled bulk foil resistors, with tolerance 1%. Absolute tolerance is not so important
in this case, as it's software calibrated anyway, but stability and tempco is what needed most. And there are only
few resistors on market which are better than these.

Resistors here used to set precise currents for sourcing thru unknown resistance for measurement. Switching is done
by MAX326's nearby. All heavily guarded and carefully designed precision voodoo.



There is also lonely Vishay PG hermetic oil-filled resistor network with tolerance 0.02% and 500.00 + 24500 resistances in it, likely setting gain for buffer amplifier (which is LT1007ACN8
in DIP8 package nearby). Switching of different gains done by DG444 and DG408 switch IC's nearby.







Switching capacitor dividers/multipliers, using film KEMET capacitors and Linear LTC1043CSW blocks. Keithley 2001 meter have only
one LTC1043 in DIP package.



Blob in center look like corrosion/salt, and I actually cleaned it carefully, but that was waste of time.
Why?
Because it is from insulator sponge, placed on protective/shielding/thermal bottom cover.



:)

More bottom side components.



0.91 ohm current sensing resistor shunt, unlike 2001, here its populated on bottom side of PCB. On 2001 it was right under 4-wire
resistor.



Those smudges/fingerprints seem to be on soldermask, as they are not erasable with alcohol. Looks dodgy, but it seem
not affecting performance..

Dual guard traces around sensitive high-impedance resistance measurement circuitry. Beauty!



Overview of bottom side of analog board, without cover.





Back cover/shield is same as 2001's, and actually bearing 2001-325 Rev.B tag on it.



Analog board is labelled 2002-102-04E-E5, designed in 1994, made in USA :)

Firmware

I have A09.

ROM, ODD
ROM, EVEN

As usual, now I have tracker and documentation project on my redmine for Keithley 2002.
Anybody welcome to download available stuff or participate in related endavours.

That's it for now..

Feel free to ask any suggestions, requests what can do/test with this meter.
I don't plan to do official calibration for it (as it's double price of 2001) at this moment, and
it seem to be pretty spot on already, at least it matches both Kei2400 and Kei2001 (which are still
only 3 weeks period from adjustment/cal) real close.

I do plan to work later on recovering schematics of Model 2002 as well, but have to finish 2001's one prior to that.

Want to try morph some functions from 2002 into 2001 (NPLC range, LTZ1000 reference, 8.5 digit native mode), as
already digital side of both meter's is same.

If anybody around have Model 2002 too, and can open it and remove analog board cover for internal photos - such
work greatly appreciated!
I want collect photo of different revisions/years for 2001/2002 series, but not going to
buy every single meter out there. Only 2002 photos found online were some old unit  (1997 datecode) on bbs.38hot.net.
Obviously not much owners tear down 6K$ MSRP DMM's, so that's why...

BR, TiN :)

P.S. Video is uploading, it will take couple hours, internet is not good here when need upload 2GB of FullHD video to tube  :phew:
« Last Edit: April 24, 2015, 01:23:02 pm by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 
The following users thanked this post: zlymex, julian1

Offline dr.diesel

  • Super Contributor
  • ***
  • Posts: 2214
  • Country: us
  • Cramming the magic smoke back in...
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #1 on: February 18, 2014, 08:18:04 pm »
Wow, excellent post my friend!    :-+    :-+    :clap:

Offline don.r

  • Frequent Contributor
  • **
  • Posts: 740
  • Country: ca
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #2 on: February 18, 2014, 08:20:37 pm »
OMG! Thanks so much for a post that will take hours of my time to view... and drool over.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #3 on: February 18, 2014, 08:38:00 pm »
Thanks for the pics, who made the 4 wire kelvin sense resistor?
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline Macbeth

  • Super Contributor
  • ***
  • Posts: 2571
  • Country: gb
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #4 on: February 18, 2014, 08:51:51 pm »
Why is the video not available (in the UK at least?)

ETA: I just checked in youtube and it says the video is being processed and come back in a few minutes. Whatever that means, I look forward to it :)

8.5 digits!? insane.
« Last Edit: February 18, 2014, 08:53:50 pm by Macbeth »
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #5 on: February 19, 2014, 12:40:38 am »
Very nice video and post TiN :clap: :-+

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #6 on: February 19, 2014, 07:21:50 am »
Nice pictures.
Keithley also used the LTZ1000 - A version, and I somewhere found the hint, that it also runs on about 90°C, perhaps indirectly from the datasheet. It's as unstable per year as the 3458A...

Would you mind identifying the divider ratio of the LTZ1000A circuitry, for determining its temperature stabilization?

Maybe that's R268, the blue metal foil (?) divider beneath the 120 Ohm resistor, or the blue resistor (?? R265??) .

As those parts seems to be custom specific, the inscript may not tell the content, so please measure the resistances, please.

Thanks.

Datasheet: operating temperature is 0..50°C, similar to HP 3458A, therefore, also around 90°C required.
I meant R268, sorry, made a typo.
Looking for a 1k / 13...15k divider..
Frank
« Last Edit: February 19, 2014, 07:44:09 am by Dr. Frank »
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #7 on: February 19, 2014, 07:34:04 am »
R285 is 3K, 12K, 0.1%, 100MW, THINFILM, according to repair manual BOM
Other network R268 is 13.1K, 13.3K, 0.1%, 100MW, THINFILM.
PTF56's are 68.1K, 0.1%, 1/8W METAL FILM, and 120 ohm is .1%, .25W, PRECISION WIREWOUND
BJT is PXT3904.

I'll recover full schematics of 2002 later, as still have 2001 to finish up.  :-//
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline plesa

  • Frequent Contributor
  • **
  • Posts: 965
  • Country: se
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #8 on: February 19, 2014, 07:44:39 pm »
Nice pictures.
Keithley also used the LTZ1000 - A version, and I somewhere found the hint, that it also runs on about 90°C, perhaps indirectly from the datasheet. It's as unstable per year as the 3458A...

Would you mind identifying the divider ratio of the LTZ1000A circuitry, for determining its temperature stabilization?

Maybe that's R268, the blue metal foil (?) divider beneath the 120 Ohm resistor, or the blue resistor (?? R265??) .

As those parts seems to be custom specific, the inscript may not tell the content, so please measure the resistances, please.

Thanks.

Datasheet: operating temperature is 0..50°C, similar to HP 3458A, therefore, also around 90°C required.
I meant R268, sorry, made a typo.
Looking for a 1k / 13...15k divider..
Frank

The resistance for temperature controll seems to be 80k.
 

Offline quarks

  • Frequent Contributor
  • **
  • Posts: 874
  • Country: de
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #9 on: February 19, 2014, 08:03:12 pm »
Great post, thanks for sharing :-+
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #10 on: February 19, 2014, 11:49:50 pm »
Very nice post! Thank you for taking the time to put that together. :)
 

Offline retiredcaps

  • Super Contributor
  • ***
  • Posts: 3575
  • Country: ca
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #11 on: February 20, 2014, 12:11:46 am »
Thank you for documenting another teardown.   :-+

It is unlikely I will ever have one of these so this is the next best thing.
 

Offline skipjackrc4

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #12 on: February 20, 2014, 12:19:27 am »
Thank you so much for that!  That thing sure looks nice.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #13 on: February 20, 2014, 12:50:21 pm »
Great work!  Thanks for sharing.
Best Wishes,

 Saturation
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #14 on: February 20, 2014, 01:05:53 pm »
One of neat features of 2002 - user accessible internal relative temperature. Calibration procedure include storing actual internal temperature, and then difference
can be shown any time.

I had +24.5°C in room, powered on cold meter, it show -22°C (minus means that current internal temperature is 22 lower, than one which was during calibration).
After 5 minutes idle running temperature difference raised quickly to -17c..



..after 38 minutes



Now meter was measuring LM399 output, logging readings every 20 seconds, temperature difference reached -4.8...-4.83 °C

This could be useful for volt-nutting, to strive to last ppms of absolute accuracy :)
« Last Edit: February 20, 2014, 01:11:39 pm by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #15 on: February 23, 2014, 09:57:00 am »
Captured some thermal shots with Fluke Ti32 today.

LTZ1000A, had meter running a hour to warmup, then removed cover and captured shots (while still running).
Square thing on top of LTZ - is sticked piece of tape, to get surface temperature. (Because shiny metal on thermacam is just a reflection of nearest objects, not temperature of surface directly).

So surface temp of LTZ package is around +45°C. This is not die temp, as we have 400°C/W ?jc per datasheet.
Ambient was +26.2°C



More to come...
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Online tszaboo

  • Super Contributor
  • ***
  • Posts: 7305
  • Country: nl
  • Current job: ATEX product design
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #16 on: February 23, 2014, 11:02:30 am »
Subscribed to the thread. I'll check out carefully in my working hours though. Thank you for sharing!
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #17 on: February 23, 2014, 04:39:44 pm »
TiN,

would you mind measuring the BE voltage of the temperature sensing transistor Q2 of the LTZ1000A, to determine its stabilization temperature? That's pin 6 -7.

Thanks.
Frank
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #18 on: February 23, 2014, 06:24:46 pm »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline Noise Floor

  • Regular Contributor
  • *
  • Posts: 117
  • Country: us
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #19 on: February 24, 2014, 01:42:18 am »
Great teardown, thanks.  :clap:
 

Offline riktw

  • Contributor
  • Posts: 12
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #20 on: February 24, 2014, 11:58:06 am »
Amazing teardown and very nice quality pictures.
I now want a DMM like that even more  :-[
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #21 on: February 27, 2014, 07:07:27 pm »
Reversed schematics of ADC module of 2002 ... Pretty big difference compared to K2001.


PNG-pic, 1000KB

PDF

Any idea how it works?  ;)
« Last Edit: July 22, 2014, 04:23:11 pm by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline uoficowboy

  • Regular Contributor
  • *
  • Posts: 186
  • Country: us
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #22 on: February 27, 2014, 07:17:11 pm »
Reversed schematics of ADC module of 2002 ... Pretty big difference compared to K2001.

PNG-pic, 500KB
PDF

Any idea how it works?  ;)
Slightly off topic - but what is your technique for getting these schematics? I've been measuring out a board recently and would love any suggestions that might speed things up :)
 

Offline TiNTopic starter

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #23 on: February 27, 2014, 07:49:23 pm »
No any special technique. Just a trusty Fluke 87V in ohm mode with buzzer (it's much quicker response and not so much test voltage as in diode mode)
and one probe on component pad, and then poke every component on board, to trace where are connections, and draw schematics in ECAD.
Plus some common sense, like first i trace all power rails, decoupling, then connector nets, then digital stuff, then go to analog. Also I combine top and bottom side photos in photoshop,
and paint out already verified components, so would reduce hassle and not test same stuff over and over (as I do this usually few hours a day, sometimes with
weeks delays in between, so might forget what already traced).

Sometimes on bigger analog board i use Keithley 2400 in voltage source mode, output 1V, 1uA, set reference fail level to 0.01V, and trace components.
I have buzzer on digital output from K2400, so i can trace copper conductors only, not worrying about opening diodes or resistors like 2 ohm, etc.


YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline anotherlin

  • Regular Contributor
  • *
  • Posts: 244
  • Country: fr
Re: Keithley 2002 8.5 digit DMM review and teardown
« Reply #24 on: February 28, 2014, 04:46:50 pm »
Reversed schematics of ADC module of 2002 ... Pretty big difference compared to K2001.

PNG-pic, 500KB
PDF

Any idea how it works?  ;)

Probably an integrating ADC ?

You mentionned you used "ECAD" for schematic capture, can you give a link to the website of this tool, please ?
"Lots of people have made $100K or more mistakes and didn't get the boot. It's called training, why fire them after such an expensive lesson?" -- EEVblog Electronics Community Forum
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf